Legal claims defining the scope of protection. Each claim is shown in both the original legal language and a plain English translation.
1. A photonic processing system comprising: an optical encoder configured to encode an input vector into a first plurality of optical signals; a photonic processor configured to: receive the first plurality of optical signals, each of the first plurality of signals received by a respective input spatial mode of a plurality of input spatial modes of the photonic processor; perform a plurality of operations on the first plurality of optical signals, the plurality of operations implementing a matrix multiplication of the input vector by a matrix; and output a second plurality of optical signals representing an output vector, each of the second plurality of signals transmitted by a respective output spatial mode of a plurality of output spatial modes of the photonic processor; and an optical receiver configured to detect the second plurality of optical signals and output an electrical digital representation of the output vector, wherein the optical encoder is configured to: encode an absolute value of a vector component of the input vector into an amplitude of a respective optical signal of the first plurality of optical signals; and encode a phase of the vector component of the input vector into a phase of the respective optical signal of the first plurality of optical signals.
The photonic processing system performs high-speed matrix multiplication using optical signals for applications in computing and signal processing. The system addresses the limitations of electronic processors in handling large-scale matrix operations, which are computationally intensive and energy-consuming. The system includes an optical encoder, a photonic processor, and an optical receiver. The optical encoder converts an input vector into a plurality of optical signals, where each signal's amplitude represents the absolute value of a vector component and its phase represents the phase of that component. The photonic processor receives these optical signals through multiple input spatial modes, performs matrix multiplication by applying a series of operations, and outputs a second set of optical signals representing the resulting vector. Each output signal is transmitted through a distinct output spatial mode. The optical receiver detects these output signals and converts them into an electrical digital representation of the output vector. The system leverages optical parallelism and high-speed signal processing to accelerate matrix computations, offering advantages in speed and energy efficiency over traditional electronic methods.
2. The photonic processing device of claim 1 , wherein the optical receiver is configured to detect the second plurality of optical signals using phase sensitive detectors.
The invention relates to photonic processing devices designed for high-speed optical signal processing. The device addresses the challenge of efficiently detecting and processing multiple optical signals in communication systems, particularly where phase information is critical for accurate data extraction. The device includes an optical receiver configured to detect a second set of optical signals using phase-sensitive detectors. These detectors are capable of resolving both the amplitude and phase of incoming optical signals, enabling advanced modulation formats and improved signal integrity. The phase-sensitive detection mechanism enhances the device's ability to distinguish between closely spaced optical signals, reducing errors in high-density optical networks. This configuration is particularly useful in coherent optical communication systems, where phase information is essential for demodulating complex modulation schemes such as quadrature phase-shift keying (QPSK) or higher-order formats. The device may also include additional components, such as optical filters or modulators, to further refine signal processing. By leveraging phase-sensitive detection, the invention improves the performance of optical communication systems, enabling higher data rates and greater reliability in signal transmission.
3. The photonic processing device of claim 2 , further comprising a light source configured to: provide first light to the optical encoder for use in encoding the first plurality of optical signals; and provide second light to the optical receiver for use as a local oscillator by the phase sensitive detectors, wherein: the local oscillator is phase coherent with each of the first plurality of optical signals; and a first path length of the first plurality of optical signals from the light source to the optical receiver is substantially equal to a second path length of the local oscillator from the light source to the optical receiver.
This invention relates to photonic processing devices, specifically those used for encoding and detecting optical signals with precise phase coherence. The device addresses the challenge of maintaining phase coherence between encoded optical signals and a local oscillator in phase-sensitive detection systems, which is critical for accurate signal processing in applications like optical communications and sensing. The photonic processing device includes an optical encoder that encodes a first plurality of optical signals, an optical receiver with phase-sensitive detectors for detecting the encoded signals, and a light source that provides two distinct light paths. The light source generates first light directed to the optical encoder for encoding the optical signals and second light directed to the optical receiver as a local oscillator. The local oscillator is phase coherent with each of the encoded optical signals, ensuring accurate phase-sensitive detection. Additionally, the device ensures that the path length of the encoded signals from the light source to the optical receiver is substantially equal to the path length of the local oscillator, minimizing phase differences and improving detection accuracy. This design enhances signal integrity and reliability in high-precision optical systems.
4. The photonic processing device of claim 1 , wherein the matrix is an arbitrary unitary matrix.
A photonic processing device processes optical signals using a matrix-based approach to perform linear transformations. The device includes an array of optical elements configured to implement a matrix operation on input optical signals, where the matrix is an arbitrary unitary matrix. Unitary matrices preserve the norm of input vectors, ensuring energy conservation in optical systems. This allows the device to perform lossless transformations, which is critical for maintaining signal integrity in photonic circuits. The arbitrary nature of the unitary matrix enables the device to handle a wide range of linear operations, including beamforming, mode conversion, and signal routing, without requiring specific constraints on the matrix structure. The optical elements may include phase shifters, beam splitters, or other components that manipulate the amplitude and phase of optical signals to achieve the desired matrix transformation. This design is particularly useful in applications such as optical computing, quantum information processing, and high-speed communication systems, where efficient and flexible signal manipulation is essential. The use of a unitary matrix ensures that the transformations are reversible and energy-efficient, addressing the challenge of signal loss in photonic systems.
5. The photonic processing device of claim 1 , further comprising: a plurality of frontends, wherein each of the plurality of frontends is associated with one input spatial mode of the plurality of input spatial modes of the photonic processor, wherein each of the plurality of frontends comprises: a plurality of optical encoders, each of the optical encoders configured to encode a respective component of an input vector into an optical signal, wherein each optical encoder is configured to output an optical signals of a wavelength different from wavelengths output by the other optical encoders; and an input wavelength division multiplexer (WDM) configured to receive each of the optical signals from each of the plurality of optical encoders in a separate spatial mode and output each of the optical signals in a single spatial mode connected to a respective input spatial mode of the plurality of input spatial modes of the photonic processor; and a plurality of backends, wherein each of the plurality of backends is associated with one output spatial mode of a plurality of output spatial modes of the photonic processor, wherein each of the plurality of backends comprises: an output wavelength division multiplexer (WDM) configured to receive optical signals of different wavelengths from a respective one of the plurality of output spatial modes of the photonic processor and output each of the optical signals of different wavelengths in a respective spatial mode of a plurality of output spatial modes of the WDM; and a plurality of optical receivers, each of the optical receivers configured to determine a respective component of an output vector by detecting a respective optical signal associated with a respective output spatial mode of the WDM.
A photonic processing device processes input data using optical signals to perform computations. The device addresses challenges in high-speed, parallel data processing by leveraging optical encoding and multiplexing to handle multiple spatial modes efficiently. The device includes multiple frontends, each associated with an input spatial mode. Each frontend contains optical encoders that convert different components of an input vector into optical signals, each at a distinct wavelength. These signals are then combined into a single spatial mode using an input wavelength division multiplexer (WDM) before being fed into the photonic processor. Similarly, the device includes multiple backends, each linked to an output spatial mode of the photonic processor. Each backend uses an output WDM to separate optical signals of different wavelengths into distinct spatial modes, which are then detected by optical receivers to reconstruct the output vector components. This architecture enables parallel processing of multiple spatial modes, enhancing computational throughput and efficiency in photonic systems.
6. The photonic processing device of claim 1 , wherein the photonic processor comprises: a first array of interconnected variable beam splitters (VBSs) comprising a first plurality of optical inputs corresponding to the first plurality of input spatial modes and a first plurality of optical outputs; a second array of interconnected VBSs comprising a second plurality of optical inputs and a second plurality of optical outputs corresponding to the plurality of output spatial modes; and a plurality of controllable optical elements, each of the plurality of these controllable optical elements coupling a single one of the first plurality of optical outputs of the first array to a respective single one of the second plurality of optical inputs of the second array.
This invention relates to photonic processing devices designed for manipulating optical signals in spatial modes. The device addresses the challenge of efficiently routing and transforming optical signals between different spatial modes, which is critical for applications in optical computing, quantum information processing, and high-speed communication systems. The photonic processor includes two interconnected arrays of variable beam splitters (VBSs). The first array receives a plurality of input spatial modes through its optical inputs and processes these signals, producing a first set of optical outputs. The second array, similarly structured, receives these outputs and transforms them into a desired set of output spatial modes. Controllable optical elements, such as phase shifters or optical switches, are used to couple each output from the first array to a specific input of the second array, enabling precise control over signal routing and transformation. The interconnected VBS arrays allow for dynamic reconfiguration of the optical signals, enabling adaptive processing based on varying input conditions. The controllable optical elements ensure that each signal path can be independently adjusted, providing flexibility in signal manipulation. This architecture enhances the device's ability to handle complex optical transformations while maintaining high efficiency and scalability. The invention is particularly useful in applications requiring real-time adaptation of optical signal paths, such as in reconfigurable optical networks or quantum computing systems.
7. The photonic processing device of claim 6 , further comprising a controller configured to: perform a singular value decomposition (SVD) of the matrix to determine a first, second, and third SVD matrix; control the first plurality of interconnected VBSs to implement the first SVD matrix; control the second plurality of interconnected VBSs to implement the second SVD matrix; and control the plurality of controllable optical elements to implement the third SVD matrix, wherein the third SVD matrix is a diagonal matrix.
A photonic processing device processes optical signals using interconnected variable beam splitters (VBSs) and controllable optical elements to perform matrix operations. The device addresses the challenge of efficiently implementing complex linear algebra operations, such as singular value decomposition (SVD), in optical systems, which is critical for applications like signal processing, machine learning, and communications. The device includes multiple interconnected VBSs and controllable optical elements, such as phase shifters or attenuators, arranged to manipulate optical signals. A controller performs SVD on an input matrix to decompose it into three matrices: the first and second matrices are implemented by the first and second sets of interconnected VBSs, respectively, while the third matrix, which is diagonal, is implemented by the controllable optical elements. This configuration allows the device to perform SVD operations entirely in the optical domain, enabling high-speed, low-power processing of optical signals. The interconnected VBSs and optical elements are configured to adjust beam splitting ratios and phase shifts to realize the matrix operations, providing a scalable and reconfigurable solution for photonic computing.
8. The photonic processing device of claim 7 , wherein the controller further comprises at least one digital-to-analog converter (DAC) to adjust one or more parameters of the first plurality of interconnected VBSs and the second plurality of interconnected VBSs.
A photonic processing device includes a controller with at least one digital-to-analog converter (DAC) to adjust parameters of interconnected variable beam splitters (VBSs). The device processes optical signals using two sets of interconnected VBSs, where the first set manipulates input signals and the second set combines or routes them. The controller dynamically configures the VBSs to control signal routing, splitting ratios, or phase adjustments, enabling flexible optical signal processing. The DAC converts digital control signals into analog adjustments for precise tuning of the VBS parameters, such as splitting ratios or phase shifts. This allows real-time reconfiguration of the optical network for applications like signal routing, modulation, or beamforming in photonic circuits. The system enhances adaptability and performance in optical communication, computing, or sensing systems by dynamically adjusting optical signal paths and properties.
9. The photonic processing device of claim 8 , wherein: each of the VBSs of the first plurality of interconnected VBSs and each of the VBSs of the second plurality of interconnected VBSs is associated with a respective address; and the at least one DAC includes a single DAC that controls a plurality of the VBSs of the first and/or second plurality of interconnected VBSs using the addresses.
A photonic processing device includes a plurality of interconnected variable beam splitters (VBSs) arranged in a network to process optical signals. The device addresses a challenge in photonic systems where precise control of optical signal routing and splitting is required for high-speed, low-power data processing. The VBSs dynamically adjust the splitting ratio of optical signals, enabling flexible signal routing and modulation. The device includes at least one digital-to-analog converter (DAC) that controls multiple VBSs by assigning each VBS a unique address. The DAC uses these addresses to selectively adjust the splitting ratios of the VBSs, allowing for precise and coordinated control of optical signal paths. This address-based control simplifies the system architecture by reducing the number of DACs needed, improving scalability and efficiency. The interconnected VBSs form two distinct groups, each with its own set of addresses, enabling independent or coordinated control of optical signals within different sections of the network. The device is particularly useful in applications requiring reconfigurable optical circuits, such as optical computing, signal processing, and telecommunications.
10. A photonic processing system comprising: an optical encoder configured to encode an input vector into a first plurality of optical signals; a photonic processor configured to: receive the first plurality of optical signals, each of the first plurality of signals received by a respective input spatial mode of a plurality of input spatial modes of the photonic processor; perform a plurality of operations on the first plurality of optical signals, the plurality of operations implementing a matrix multiplication of the input vector by a matrix; and output a second plurality of optical signals representing an output vector, each of the second plurality of signals transmitted by a respective output spatial mode of a plurality of output spatial modes of the photonic processor; and an optical receiver configured to detect the second plurality of optical signals and output an electrical digital representation of the output vector; wherein the matrix is a first matrix and the photonic processing device further comprises a controller configured to control the photonic processing device to perform multiplication of a second matrix by the first matrix by: (a) determining a plurality of input vectors from each column of the second matrix; (b) selecting an input vector from the plurality of input vectors; (c) encoding the selected input vector into the first plurality of optical signals using the optical encoder; (d) performing the plurality of operations on the first plurality of optical signals associated with the first input vector; (e) detecting the second plurality of optical signals associated with the selected input vector; (f) storing digital detection results based on the detected second plurality of optical signals; (g) repeating acts (b)-(f) for the other input vectors of the plurality of input vectors; and (h) digitally combine the digital detection results to determine a resulting matrix resulting from the multiplication of the second matrix by the first matrix.
A photonic processing system performs matrix multiplication using optical signals for high-speed computation. The system addresses the challenge of efficiently implementing large-scale matrix operations, which are computationally intensive in conventional electronic systems. The system includes an optical encoder that converts an input vector into multiple optical signals, each corresponding to a spatial mode. A photonic processor receives these signals, performs matrix multiplication by applying operations to the optical signals, and outputs a new set of optical signals representing the result. An optical receiver converts these signals into a digital electrical output. The system can also perform sequential matrix multiplications by decomposing a second matrix into column vectors, processing each vector individually, and combining the results digitally. This approach leverages optical parallelism for faster computation while using digital processing for flexibility in handling complex operations. The system is particularly useful in applications requiring real-time processing of large matrices, such as machine learning and signal processing.
11. The photonic processing device of claim 1 , wherein the optical receiver comprises a low-pass filter configured to perform an analog summation of multiple subsequent signals associated with each output spatial mode of the plurality of output spatial modes of the photonic processor.
A photonic processing device includes an optical receiver with a low-pass filter designed to perform an analog summation of multiple subsequent signals associated with each output spatial mode of a photonic processor. The photonic processor generates a plurality of output spatial modes, each carrying distinct signal components. The low-pass filter in the optical receiver integrates these signals over time, effectively summing them to produce a combined output. This summation helps mitigate noise and enhance signal integrity by averaging fluctuations across multiple signal instances. The device leverages spatial mode diversity to improve processing robustness, particularly in applications requiring high-speed or high-fidelity signal handling. The low-pass filter ensures that transient variations are smoothed, providing a stable output. This approach is useful in optical communication systems, signal processing, and computing applications where maintaining signal coherence and reducing noise are critical. The integration of the low-pass filter directly within the optical receiver streamlines the processing pipeline, eliminating the need for additional external components. The system is particularly advantageous in environments where real-time signal aggregation is required, such as in photonic neural networks or high-speed data transmission systems.
12. A method of optically performing matrix-vector multiplication, the method comprising: receiving a digital representation of an input vector; encoding, using an optical encoder, the input vector into a first plurality of optical signals; performing, using a processor, a singular value decomposition (SVD) of a matrix to determine a first, second, and third SVD matrix; controlling a photonic processor comprising a plurality of variable beam splitters (VBS) to optically implement the first, second, and third SVD matrix; propagating the first plurality of optical signals through the photonic processor; detecting a second plurality of optical signals received from the photonic processor; and determining an output vector based on the detected second plurality of optical signals, wherein the output vector represents a result of the matrix-vector multiplication, wherein encoding the input vector comprises: encoding an absolute value of a vector component of the input vector into an amplitude of a respective optical signal of the first plurality of optical signals; and encoding a phase of the vector component of the input vector into a phase of the respective optical signal of the first plurality of optical signals.
This invention relates to optical computing, specifically a method for performing matrix-vector multiplication using photonic processors to accelerate computations. The problem addressed is the computational inefficiency of traditional electronic matrix-vector multiplication, particularly for large-scale applications like machine learning and signal processing. The solution involves an optical approach that leverages singular value decomposition (SVD) and photonic hardware to enhance speed and energy efficiency. The method begins by receiving a digital input vector and encoding it into optical signals using an optical encoder. Each component of the input vector is encoded into both the amplitude and phase of an optical signal, representing its absolute value and phase, respectively. A processor then performs SVD on a target matrix to decompose it into three matrices (U, Σ, V^T). A photonic processor, equipped with variable beam splitters (VBS), is configured to optically implement these SVD matrices. The encoded optical signals propagate through the photonic processor, where they undergo transformations corresponding to the matrix operations. The resulting optical signals are detected, and the output vector is determined from these signals, representing the product of the matrix and input vector. This approach combines digital preprocessing with optical computation, enabling high-speed matrix-vector multiplication while maintaining accuracy. The use of SVD decomposition allows for efficient optical implementation of complex matrix operations, reducing reliance on electronic processing.
13. The method of claim 12 , wherein the detecting of the second plurality of optical signals is performed using phase-sensitive detectors.
This invention relates to optical signal detection systems, specifically improving the accuracy and sensitivity of detecting optical signals in applications such as telecommunications, sensing, or imaging. The problem addressed is the need for more precise detection of optical signals, particularly when distinguishing between multiple signals or extracting phase information, which is critical for high-performance optical systems. The method involves detecting a second plurality of optical signals using phase-sensitive detectors. These detectors are designed to measure not only the intensity of the optical signals but also their phase, allowing for enhanced discrimination between signals and improved signal-to-noise ratio. The phase-sensitive detection is particularly useful in systems where phase information is critical, such as in coherent optical communication, interferometry, or quantum optics. The method builds upon a broader approach that includes generating and modulating optical signals, transmitting them through a medium, and detecting them at a receiver. The use of phase-sensitive detectors ensures that subtle phase variations in the optical signals are accurately captured, which is essential for applications requiring high precision, such as fiber-optic communications, lidar systems, or optical sensing. By incorporating phase-sensitive detection, the system achieves better performance in noisy environments and improves the overall reliability of optical signal processing.
14. The method of claim 13 , further comprising: providing first light, from a light source, to the optical encoder for encoding the first plurality of optical signals; and providing second light, from the light source, to the phase-sensitive detectors, wherein the second light is used as a local oscillator by the phase-sensitive detectors.
This invention relates to optical encoding and detection systems, specifically addressing challenges in accurately encoding and decoding optical signals while minimizing noise and interference. The system includes an optical encoder that processes a first plurality of optical signals, converting them into encoded optical signals. These encoded signals are then transmitted through an optical channel, which may introduce distortions or noise. To mitigate these issues, the system employs phase-sensitive detectors that receive the encoded signals and demodulate them using a local oscillator. The local oscillator is derived from the same light source that provides illumination for the optical encoder, ensuring phase coherence and improving detection accuracy. The light source generates both the first light used for encoding and the second light used as the local oscillator, enhancing synchronization and reducing phase-related errors. This approach improves signal fidelity in optical communication systems, particularly in applications requiring high precision, such as fiber-optic networks or optical sensing systems. The invention focuses on optimizing the interaction between the encoder, optical channel, and detectors to achieve robust signal transmission and detection.
15. The method of claim 14 , wherein: the local oscillator is phase coherent with each of the first plurality of optical signals; and a first path length of the first plurality of optical signals from the light source to the phase-sensitive detectors is substantially equal to a second path length of the local oscillator from the light source to the phase-sensitive detectors.
This invention relates to optical signal processing, specifically improving phase coherence and path length matching in optical communication or sensing systems. The problem addressed is maintaining precise phase alignment between a local oscillator and multiple optical signals in systems where path length differences can introduce phase errors, degrading performance. The method involves generating a first plurality of optical signals from a light source and a local oscillator signal, also derived from the light source. The local oscillator is phase coherent with each of the optical signals, ensuring synchronized phase relationships. The optical signals and the local oscillator are routed to phase-sensitive detectors, with the path length from the light source to the detectors for the optical signals being substantially equal to the path length for the local oscillator. This path length matching minimizes phase discrepancies caused by optical path differences, improving detection accuracy and system reliability. The technique is particularly useful in applications requiring high-precision phase detection, such as coherent optical communication, interferometry, or quantum sensing, where phase coherence and path length control are critical for performance. By ensuring both phase coherence and equal path lengths, the method enhances signal fidelity and reduces errors in phase-sensitive measurements.
16. The method of claim 12 , wherein the matrix is an arbitrary unitary matrix.
The invention relates to quantum computing and error correction, specifically addressing the challenge of implementing fault-tolerant quantum operations using arbitrary unitary matrices. In quantum computing, maintaining coherence and correcting errors is critical, but traditional methods often rely on specific gate decompositions that limit flexibility. This invention provides a method for encoding quantum information into a logical qubit using an arbitrary unitary matrix, enabling more versatile error correction schemes. The unitary matrix is applied to a set of physical qubits to transform their state into a logical qubit state, where the matrix's properties ensure that errors can be detected and corrected without collapsing the quantum information. The method involves selecting an arbitrary unitary matrix, applying it to the physical qubits, and then performing measurements to identify and correct errors while preserving the logical qubit state. This approach allows for the use of custom unitary matrices tailored to specific error correction requirements, improving fault tolerance in quantum computations. The invention also includes steps for verifying the correctness of the error correction process, ensuring robustness against various types of quantum noise. By leveraging arbitrary unitary matrices, the method enhances the adaptability and reliability of quantum error correction protocols.
17. The method of claim 12 , further comprising performing multiple matrix-vector multiplications simultaneously using wavelength division multiplexing.
This invention relates to high-speed data processing systems, specifically methods for accelerating matrix-vector multiplications in computational tasks. The core problem addressed is the computational bottleneck in performing large-scale matrix-vector multiplications, which are fundamental operations in machine learning, scientific computing, and signal processing. Traditional approaches rely on sequential processing or parallelization through electronic hardware, which may not scale efficiently for high-dimensional data. The invention describes a method that enhances computational efficiency by performing multiple matrix-vector multiplications simultaneously using wavelength division multiplexing (WDM). WDM is a technique traditionally used in optical communications to transmit multiple data streams over a single optical fiber by assigning each stream a distinct wavelength. In this method, WDM is adapted to parallelize matrix-vector operations by encoding different matrix-vector pairs onto different wavelengths of light. This allows multiple computations to occur concurrently within a single optical processing unit, significantly increasing throughput compared to electronic-only solutions. The method involves encoding input vectors into optical signals at different wavelengths, processing these signals through an optical system configured to perform matrix-vector multiplications, and then decoding the resulting outputs. The optical system may include components such as optical modulators, beam splitters, and detectors that enable parallel processing. By leveraging WDM, the system avoids the latency and power constraints of electronic parallelization, making it suitable for real-time applications requiring high-speed computations. The approach is particularly advanta
18. The method of claim 17 , wherein: the input vector is one of a plurality of input vectors: the method further comprises encoding each of the plurality of input vectors into a respective one of a first plurality of optical signal of a particular wavelength, wherein each wavelength associated with each one of the first plurality of optical signals is different from the other wavelengths of the other ones of the first plurality of optical signals.
This invention relates to optical signal processing, specifically a method for encoding multiple input vectors into distinct optical signals for parallel processing. The problem addressed is the need to efficiently encode and transmit multiple data vectors simultaneously using optical signals, ensuring each vector is uniquely identifiable and processed without interference. The method involves receiving a plurality of input vectors, where each vector represents a distinct set of data. Each input vector is encoded into a corresponding optical signal, with each optical signal having a unique wavelength. The wavelengths are selected such that they are distinct from one another, preventing overlap or interference between the encoded signals. This allows multiple optical signals to be transmitted and processed in parallel, leveraging the high bandwidth and low latency of optical communication systems. The use of unique wavelengths for each encoded signal ensures that the data remains distinguishable during transmission and processing. This approach is particularly useful in applications requiring high-speed, parallel data processing, such as optical computing, telecommunications, and data center networks. The method enables efficient multiplexing of multiple data streams, improving throughput and reducing latency in optical communication systems.
19. The method of claim 12 , wherein the photonic processor comprises: a first array of interconnected variable beam splitters (VBSs) comprising a first plurality of optical inputs corresponding to the first plurality of input spatial modes and a first plurality of optical outputs; a second array of interconnected VBSs comprising a second plurality of optical inputs and a second plurality of optical outputs corresponding to the plurality of output spatial modes; and a plurality of controllable optical elements, each of the plurality of these controllable optical elements coupling a single one of the first plurality of optical outputs of the first array to a respective single one of the second plurality of optical inputs of the second array.
This invention relates to photonic processors designed for manipulating optical signals in multiple spatial modes. The problem addressed is the need for efficient and programmable control of optical signal routing and transformation in photonic circuits, particularly for applications in quantum computing, optical communications, and signal processing. The photonic processor includes two interconnected arrays of variable beam splitters (VBSs). The first array receives a plurality of input spatial modes and splits or combines optical signals across its interconnected VBSs, producing a first set of optical outputs. The second array, similarly structured, processes these outputs to generate a plurality of output spatial modes. Controllable optical elements, such as phase shifters or modulators, are placed between the two arrays, each coupling a single output from the first array to a single input of the second array. These elements enable precise control over the optical signals, allowing for programmable reconfiguration of the photonic circuit. The interconnected VBSs in each array allow for dynamic adjustment of signal distribution, while the controllable optical elements between the arrays provide fine-grained control over phase, amplitude, or other properties of the optical signals. This architecture enables flexible and scalable manipulation of optical signals in multiple spatial modes, supporting advanced photonic applications.
20. The method of claim 19 , wherein: the first plurality of interconnected VBSs implements the first SVD matrix; the second plurality of interconnected VBSs implements the second SVD matrix; and the plurality of controllable optical elements implements the third SVD matrix, wherein the third SVD matrix is a diagonal matrix.
This invention relates to optical computing systems that use interconnected vector beam splitters (VBSs) and controllable optical elements to perform matrix operations, specifically singular value decomposition (SVD). The problem addressed is the need for efficient, scalable, and reconfigurable optical systems capable of implementing large-scale matrix decompositions, which are computationally intensive in traditional electronic systems. The system includes multiple interconnected VBSs that collectively implement the first and second SVD matrices. These VBSs are arranged to split and combine optical signals in a manner that replicates the linear transformations defined by the SVD matrices. Additionally, a plurality of controllable optical elements, such as phase shifters or attenuators, implements the third SVD matrix, which is a diagonal matrix. These elements adjust the amplitude or phase of optical signals to achieve the desired diagonal transformation. The interconnected VBSs and controllable optical elements work together to perform the full SVD operation, enabling efficient decomposition of input optical signals into singular values and vectors. The system is reconfigurable, allowing adaptation to different matrix dimensions and decomposition requirements. This approach leverages optical parallelism and scalability to accelerate matrix computations, making it suitable for applications in signal processing, machine learning, and high-performance computing.
21. The method of claim 19 , wherein: each of the VBSs of the first plurality of interconnected VBSs and each of the VBSs of the second plurality of interconnected VBSs is associated with a respective address; and the VBSs of the first and/or second plurality are controlled by at least one digital to analog converter (DAC) that controls a plurality of the VBSs using the addresses.
A system and method for controlling variable beam splitters (VBSs) in an optical network involves interconnected VBSs organized into at least two groups, each group comprising multiple VBSs. Each VBS in both groups is assigned a unique address, enabling selective control. A digital-to-analog converter (DAC) manages the VBSs by addressing individual units or groups, adjusting their optical properties such as beam splitting ratios or reflection/transmission states. This approach allows dynamic reconfiguration of optical pathways within the network, optimizing light routing for applications like optical switching, signal modulation, or beam steering. The addressed control mechanism ensures precise and scalable manipulation of multiple VBSs, enhancing flexibility in optical system design. The system may be used in telecommunications, optical computing, or laser-based applications where adaptive beam control is required. The DAC-driven addressing simplifies integration with digital control systems, enabling real-time adjustments without manual intervention. The invention addresses challenges in managing complex optical networks by providing a scalable, addressable control framework for VBS arrays.
22. A method of optically performing matrix-vector multiplication, the method comprising: receiving a digital representation of an input vector; encoding, using an optical encoder, the input vector into a first plurality of optical signals; performing, using a processor, a singular value decomposition (SVD) of a matrix to determine a first, second, and third SVD matrix; controlling a photonic processor comprising a plurality of variable beam splitters (VBS) to optically implement the first, second, and third SVD matrix; propagating the first plurality of optical signals through the photonic processor; detecting a second plurality of optical signals received from the photonic processor; and determining an output vector based on the detected second plurality of optical signals, wherein the output vector represents a result of the matrix-vector multiplication; wherein the matrix-vector multiplication is one of a plurality of matrix-vector multiplications performed to perform matrix-matrix multiplication, wherein the matrix is a first matrix and the matrix-matrix multiplication comprises multiplication of a second matrix by the first matrix by, the method further comprising: (a) determining a plurality of input vectors from each column of the second matrix; (b) selecting an input vector from the plurality of input vectors; (c) encoding the selected input vector into the first plurality of optical signals using the optical encoder; (d) performing the plurality of operations on the first plurality of optical signals associated with the first input vector; (e) detecting the second plurality of optical signals associated with the selected input vector; (f) storing digital detection results based on the detected second plurality of optical signals; (g) repeating acts (b)-(f) for the other input vectors of the plurality of input vectors; and (h) digitally combine the digital detection results to determine a resulting matrix resulting from the multiplication of the second matrix by the first matrix.
This invention relates to optical computing systems for performing matrix-vector and matrix-matrix multiplications efficiently. The problem addressed is the computational complexity and energy consumption of traditional electronic matrix operations, particularly in large-scale applications like machine learning and linear algebra. The solution involves an optical approach that leverages photonic processors and singular value decomposition (SVD) to accelerate these operations. The method begins by receiving a digital input vector and encoding it into optical signals using an optical encoder. A processor performs SVD on a target matrix to decompose it into three matrices, which are then implemented optically using a photonic processor with variable beam splitters. The encoded optical signals propagate through the photonic processor, where they undergo transformations corresponding to the SVD matrices. The resulting optical signals are detected and converted into a digital output vector, representing the matrix-vector multiplication result. For matrix-matrix multiplication, the method processes each column of the second matrix as an input vector, sequentially encoding and propagating them through the photonic processor. The detected results for each input vector are stored digitally and later combined to produce the final matrix product. This approach reduces computational overhead by offloading matrix operations to optical hardware, improving speed and energy efficiency.
Unknown
September 1, 2020
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.